Skip to navigationSkip to content

'Disastrous' lack of diversity in AI industry perpetuates bias, study finds

By the Guardian

Report says an overwhelmingly white and male field has reached ‘a moment of reckoning’ over discriminatory systemsRead full story

Comments

  • Also share to
  • Unlike other historic industrial and societal revolutions (AI is both btw) we now have the awareness to catch this early before we perpetuate a trajectory of prejudice. This is critical considering the exponential speed AI is moving which means that any bias or prejudice will have very fast and far reaching effects and it's hard to tell how these will compound with AI behind the wheel more often.

  • AI's achilles heel is the problem of "edge cases", or conditions that were not featured in training data or anticipated by the people doing the training.

    When deep learning networks involve people -- classifying photos of them or their writing, for example -- these edge cases are an artifact of human diversity.

    This puts a large burden on the people training the deep learning networks to think expansively and understand why "one size fits all" approaches may disadvantage one subgroup vs another

    AI's achilles heel is the problem of "edge cases", or conditions that were not featured in training data or anticipated by the people doing the training.

    When deep learning networks involve people -- classifying photos of them or their writing, for example -- these edge cases are an artifact of human diversity.

    This puts a large burden on the people training the deep learning networks to think expansively and understand why "one size fits all" approaches may disadvantage one subgroup vs another.

    This takes creativity and empathy, the latter of which is sometimes deficient in tech workers who have lived relatively sheltered lives through elite colleges directly into isolated corporate campuses.

  • Eric Artisan
    Eric ArtisanManaging Director at Artisan Venture

    Why doesn’t tech face the intensity of scrutiny as Wall Street? Imagine if a field in finance/Wall Street ghat was comprised of 80% men. The media and government would go nuts. But I’m tech? Not so. And one could argue that tech will have a much bigger role in the future than Wall Street. It’s disturbing that80% of AI professors are men. And less than 15% of AI professionals at FB and GOOG are men. Think of all the talent we are missing out on. We need to make these opportunities and roles more accessible/appealing

    Why doesn’t tech face the intensity of scrutiny as Wall Street? Imagine if a field in finance/Wall Street ghat was comprised of 80% men. The media and government would go nuts. But I’m tech? Not so. And one could argue that tech will have a much bigger role in the future than Wall Street. It’s disturbing that80% of AI professors are men. And less than 15% of AI professionals at FB and GOOG are men. Think of all the talent we are missing out on. We need to make these opportunities and roles more accessible/appealing to females. We can’t afford to not have 50% of the population engaged and lead the path forward on AI. The costs are way too high. So, let’s focus more on such disturbing trends and call out VCs to encourage their portfolio companies to hire more female and minorities. We need diversity of ideas. Let’s encourage more universities to promote AI opportunities aggressively to minorities and women. If high standards are required in financial services, let’s apply the same rigor to tech. I hate to be a broken record but there’s too much to lose here if we don’t promote diversity.

  • Mark Douglas
    Mark DouglasproPresident & CEO at SteelHouse

    “The core of the problem is whether market forces are going to be sufficient for this to be fixed,”

    The core problem is a shortage of female and minority students pursuing computer science and other tech/science degrees. That’s not an excuse, it’s a fact. If we started focusing on guiding high school students toward tech degrees today, we could start to have this problem fixed across tech in 48 months. And we would have a leading metric to focus on rather than trailing metric.

  • Weiyee IN
    Weiyee INChief Strategy Officer

    While there is the blatantly obvious injustice that seems to need to be addressed in the lack of diversity and inclusion for the AI industry, there is actually a bigger and longer term strategic problem (my decade long appeal on Wall Street to the baser instincts of self interest for tech companies and investors to think about diversity and inclusion). AI, especially in the United States desperately needs diversity and inclusion, not just because it is the right thing to do, but in order to have

    While there is the blatantly obvious injustice that seems to need to be addressed in the lack of diversity and inclusion for the AI industry, there is actually a bigger and longer term strategic problem (my decade long appeal on Wall Street to the baser instincts of self interest for tech companies and investors to think about diversity and inclusion). AI, especially in the United States desperately needs diversity and inclusion, not just because it is the right thing to do, but in order to have a competitive advantage at national level. The breadth and depth of technology innovation and development are severely impacted by diversity and inclusion, not only at board of directors level, but also within teams in technology companies. The competitive advantage of America as the melting pot of heterogeneity will be lost if we are not able to maintain diversity and inclusion in innovation. Even if we are motivated only by selfishness, it is in our best interest as a nation to embrace diversity and inclusion if only for her competitive advantage in innovation.

  • Tom Swan
    Tom Swan

    Yes, bias based on the code writer's belief system is further proof that AI is a marketing myth, promoted by those who believe that more processing power and memory will eventually result in "human like" intelligence.

    These myopic individuals care little about

    the study of human consciousness, which is little understood. They don't see the need to

    strive for cogent self aware machines. They are selling processors. Those processors run software that is coded by humans, mostly younger white guys, who like Cheetos.

  • Monica Liu
    Monica Liu

    It’s funny how much attention people seem to give towards issues such as AI industry lacking of diversity while in reality many, many, many industries have the exact same problem in a much worse level yet nobody gives a

Want more conversations like this?

Join the Quartz community for all the intelligence, without the noise.

App Store BadgeGoogle Play Badge
Leaderboard Screenshot

A community of leaders, subject matter experts, and curious minds bringing nuance back to how we talk about the news.

Editors' Picks Screenshot

No content overload: our editors will curate the most notable and discussion-worthy pieces for you every day.

Share Screenshot

Don’t just read the story, tell it: contribute your ideas and experience to the dialogue.